ACAtune options vs OS 3.9 CPU command

Caution: Non registered users only see threads and messages in the currently selected language, which is determined by their browser. Please create an account and log in to see all content by default. This is a limitation of the forum software.


Also users that are not logged in can not create new threads. This is a, unfortunately needed, counter measure against spam. Please create an account and log in to start new threads.

Don't Panic. Please wash hands.
  • I've just bought a used ACA 1231 68000-42 card to replace my DKB Cobra because I like the small form factor and modern speeds. I have OS 3.9 BB4 installed on my A1200. I've recently put CPU fastrom burst in my startup which seems faster. What would the best option be: to keep CPU command in startup, or use ACAtune and options?

    Thanks.

  • The "CPU fastrom" command will activate the MMU to map the Kickstart into fastmem. This will not only cost you a bit of fastmem, but also a few extra cycles for MMU table lookup.


    If you're after the last bit of performance, you should not activate the MMU, but only use the CPU command to activate cache and burst (AFAIR, DCache isn't activated by default, so the command line would be "CPU cache burst").


    For Kickstart mapping, the ACA1231 has a reserved memory space that's not in the freemem list. If you make use of it with "ACAtool maprom **" (the two stars stand for "copy all 1MB including all mirrors"), you'll have slightly more performance compared to the "CPU fastrom" option, because the MMU is not involved. The mapping is handled by the memory controller of the ACA1231, and that is so fast that it does not take any performance away (mapping happens in parallel to other operations). This has been the case since the very first ACA1230, and has been continued in all of our successor boards, no matter what CPU.

  • I suggest to not use OS3.x - we don't support that. Look at OS3.1.4 or OS3.2 instead. Further, the "CPU" command will not take advantage of the features of the ACA1230/42 - only ACAtune will. If you want to, you can activate cache and burst with the CPU command. However, the EC030 CPU does not have an MMU, so "Fastrom" won't work.


    There is no OS-specific command line for ACAtune. I'd activate MapROM (which will also work without MMU, and it's even faster), and I'd always set the Kick3.0 jumper, so $c0 memory is always active.

  • I suggest to not use OS3.x - we don't support that. Look at OS3.1.4 or OS3.2 instead.

    There seems to be similar stances from both icomp and Cloanto regarding Authenticity, and from this users point of view, it's a view I agree with. 3.1.4 and 3.2 are a big "maybe" but seem to be moving away from the concept.?(

  • I'm not discussing the legal aspect at all - I do know that Cloanto has bought the Amiga company and brand and he may have every right to sell some software and call it an OS, ROM or whatever else.


    However, the mere fact that he bought a company doesn't mean that I like or even endorse it. I believe Michael's actions are destructive, that's all. To me, the Amiga name has always kept a creative, constructive vibe. Cloanto has taken that away. Me not supporting him and his products is merely keeping up the illusion that "Amiga" is a _nice_ hobby. Please don't break it.

  • A1200 3.1 ROM OS 3.9 BB4 - rom updated (3.9) from Devs file - ROM Updates (312k) during boot.

    - ACA 1230 EC 25 @ 28.7 mhz

    "ACATUNE MAPROM * >NIL:" in startup sequence returns Maprom Disabled when ACATUNE STATUS is queried from Workbench?

    "ACATUNE MAPROM *" when no startup-sequence returns Invalid argument. ACATUNE STATUS returns No Board Found.?

    "ACATUNE MAPROM * " from Workbench gives affirmation of success 0$b0000.... - returns Maprom Activated when ACATUNE STATUS is queried (Workbench). - (this is ineffectual obviously.)

    KS 3.0 jumper is closed as you've previously suggested. Manual states leave jumper open if KS 3.1?

  • ACAATune only accepting "**" (two stars) for MapROM is known, and it's not a problem; it only copies over the 512k mirror into the already-reserved additional 512k for MapROM. It does not eat any extra mem, and it works nicely. We're leaving the bug in, as the double-star is a proper workaround with no side effects.


    KS 3.0 jumper is closed as you've previously suggested. Manual states leave jumper open if KS 3.1?

    That's just a choice, not a necessity. The 3.0 jumper activates $c0 trapdoor memory from a cold start, and this is actually the behaviour of all later ACA cards, as it does not do any harm on 3.1 and later Kickstarts.


    Jens

  • I followed this thread: ACATune - Page 1 (amiga.org)

    It worked when ACATune was in the User-Startup; but I think ACATune overrode the patched scsi.device from Setpatch and I had no IDE if soft-booted. I had to switch it off and on again at the power.

    I'm now thinking ACATune may work after Loadmodule and before Setpatch in the startup-sequence.


    SYSInfo shows 62MB fast ram. When Maprom works it's at $0bf00000. Is the KS 3.0 jumper $0C000000 memory unused??

  • I followed this thread:

    Why don't you follow this instead? It clearly says that you should put ACAtune as the first line of your startup-sequence.


    SYSInfo shows 62MB fast ram. When Maprom works it's at $0bf00000. Is the KS 3.0 jumper $0C000000 memory unused??

    $0c00.0000 memory is taken from the 64M physical fastmem block at the 62MByte boundary. This automatically stops memory recognition at 62MBytes, and you have the 1MByte at $0c00.0000, giving you a total of 63MBytes of fastmem. The 1MB that's "missing" is reserved for MapROM. It always uses 1MByte, no matter if you're mapping a 512k or a 1024k ROM, because the mirror at $00e0.0000 is also properly used. That's actually what "*" instead of "**" should do, but there seems to be a bug.


    Jens

  • Thanks Jens. I didn't know the ACATune options are case sensitive. Without the p option the IDE isn't found on reboot as it's too fast. With the p option, soft-boot is ok but... - the 3.9 OS goes into recoverable alert with every function call on WB as though the GUI is patched with the "pause".

  • It's not the GUI that is patched - only a small wait-and-do-nothing routine is executed on reset.


    This looks like OS3.9 is checking for the 3.1 ROMs and finds that the checksum is not OK. I believe I have stated often enough that I consider OS3.5 and OS3.9 to be dead forks, and we neither test, nor develop for them. We make sure everything works under 3.0, 3.1 and the new Hyperion OS versions.


    Unfortunately, the "Kicksum" option is only good for 256k Kickstarts, and I'm pretty sure that the checksum at 0xFFFFE8 is patched accordingly, otherwise you'd see a red screen on reset and the computer wouldn't start at all. So at this point, the best recommendation I can give is to use a different OS version.


    In later accelerators, I have implemented a slow-down method after reset that makes the -p option unnecessary. This was only introduced with the ACA1231, and kept in for later accelerators.


    Jens

  • Ok Thanks Jens. I have a ACA1231 but that now has a home in another A1200. I have 6 ACA cards - it's nice to know their functions. I've left the command as ACAtune -cache on -burst on >NIL: and that will have to do for this setup. The speed difference isn't a great deal and seems more perceptual from my point of view.

    Thanks for all the information and assistance.